# Defining an approach to empower clinical geneticists to do genomic reanalysis

## TL;DR

This paper presents a method to help clinical geneticists improve genetic diagnoses by reanalyzing genomic data using a decision support tool.

## Contribution

A practical approach for integrating genomic reanalysis into routine clinical follow-up using SimulConsult software.

## Key findings

- Reanalysis increased diagnostic yield from 7 to 10 (43%) for routine cases at St. George’s Hospital.
- The method identified diagnoses not previously included in variant tables, including a non-coding variant.
- Rapid reanalysis requires minimal effort and no new lab costs.

## Abstract

Sequencing reanalysis can benefit from the inclusion of new information about the patient and from the literature. We studied approaches needed to make reanalysis part of routine follow-up by clinical geneticists.

Reanalysis used the SimulConsult diagnostic decision support software, which generates a pertinence metric for gene zygosities determined from the variant table and the patient’s findings. Twenty patients had routine exome sequencing at St. George’s Hospital (London, UK). Twenty were admitted to the Undiagnosed Diseases Network at Vanderbilt University Medical Center (VUMC) and had all remained undiagnosed despite previous evaluations and sequencing.

For St. George’s cases, reanalysis picked 7 of the 7 initial diagnoses plus 2 diagnoses found later, and suggested another diagnosis with a gene absent from the variant table. For VUMC, reanalysis picked 5 of 8 diagnoses that were in the variant tables, and suggested a non-coding variant absent from the variant table.

Rapid reanalysis by clinicians could increase the yield of genetic diagnosis with minimal effort and no new lab expenses. For the routine cases at St. George’s, diagnostic yield increased from 7 to 10 (43%). Capabilities that could further increase yield include joint variant calling, robust phenotyping, clinical correlation after sequencing, and adding CNV data to variant tables.